library(tidyverse)
## ── Attaching packages ───────────────────────────────────────────────────────────── tidyverse 1.2.1 ──
## ✔ ggplot2 3.2.1 ✔ purrr 0.3.2
## ✔ tibble 2.1.3 ✔ dplyr 0.8.3
## ✔ tidyr 1.0.0 ✔ stringr 1.4.0
## ✔ readr 1.3.1 ✔ forcats 0.4.0
## ── Conflicts ──────────────────────────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
library(plotly)
##
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
##
## last_plot
## The following object is masked from 'package:stats':
##
## filter
## The following object is masked from 'package:graphics':
##
## layout
library(rvest)
## Loading required package: xml2
##
## Attaching package: 'rvest'
## The following object is masked from 'package:purrr':
##
## pluck
## The following object is masked from 'package:readr':
##
## guess_encoding
library(readr)
library(viridis)
## Loading required package: viridisLite
library(leaflet)
library(patchwork)
knitr::opts_chunk$set(
echo = TRUE,
warning = FALSE,
fig.width = 8,
fig.height = 6,
out.width = "90%"
)
options(
ggplot2.continuous.colour = "viridis",
ggplot2.continuous.fill = "viridis"
)
scale_colour_discrete = scale_colour_viridis_d
scale_fill_discrete = scale_fill_viridis_d
theme_set(theme_minimal() + theme(legend.position = "bottom"))
data_2018 =
read_csv("./data/2018data.csv") %>%
janitor::clean_names()
## Parsed with column specification:
## cols(
## .default = col_character(),
## TIME = col_time(format = ""),
## `ZIP CODE` = col_double(),
## LATITUDE = col_double(),
## LONGITUDE = col_double(),
## `NUMBER OF PERSONS INJURED` = col_double(),
## `NUMBER OF PERSONS KILLED` = col_double(),
## `NUMBER OF PEDESTRIANS INJURED` = col_double(),
## `NUMBER OF PEDESTRIANS KILLED` = col_double(),
## `NUMBER OF CYCLIST INJURED` = col_double(),
## `NUMBER OF CYCLIST KILLED` = col_double(),
## `NUMBER OF MOTORIST INJURED` = col_double(),
## `NUMBER OF MOTORIST KILLED` = col_double(),
## COLLISION_ID = col_double()
## )
## See spec(...) for full column specifications.
newnames = colnames(data_2018) %>%
str_replace("number_of_","")
names(data_2018) = newnames
tidy_data =
data_2018 %>%
mutate(
date_complete = date
) %>%
separate(date, into = c("month", "day", "year"), sep = "/") %>%
separate(time, into = c("hour", "minute"), sep = ":") %>%
select(-zip_code, -location, -on_street_name, -cross_street_name, -off_street_name,-collision_id,-year) %>%
rename("vehicle_type" = "vehicle_type_code_1") %>%
mutate( day = as.numeric(day),
month = as.numeric(month),
hour = as.numeric(hour),
minute = as.numeric(minute),
latitude = replace_na(latitude,0),
vehicle_type = str_to_lower(vehicle_type)
) %>%
filter( latitude != 0)
Vehicle type
vehicle_type_data =
tidy_data %>%
mutate(
vehicle_type = replace(vehicle_type,str_detect(vehicle_type,"truck"),"truck"),
vehicle_type = replace(vehicle_type,str_detect(vehicle_type,"sport utility"),"sport utility vehicle")
) %>%
filter( vehicle_type %in% c("taxi","passenger vehicle","truck","sport utility vehicle")) %>%
group_by(vehicle_type,hour) %>%
summarize(
n = n()
)
vehicle_type_data %>%
plot_ly(
x = ~hour, y = ~n, color = ~vehicle_type, type = "scatter", mode = "line") %>%
layout(
title = "Collisions of Day for Different Vehicles",
xaxis = list(title = "Hour of Day"),
yaxis = list(title = "Collisions")
)
Top 8 Collision Reasons
reason_data =
tidy_data %>%
group_by(contributing_factor_vehicle_1) %>%
summarize(n = n()) %>%
arrange(desc(n)) %>%
head(10)
reason_data %>%
plot_ly(x = ~reorder(contributing_factor_vehicle_1,desc(n)), y = ~n, color = ~contributing_factor_vehicle_1 ,type = "bar") %>%
layout(
title = "The Number of Items Ordered in Each Aisle",
xaxis = list(title = "Different Reasons"),
yaxis = list(title = "Count")
)
Mapping
data_2018 = tidy_data
data_2018 = rename(data_2018, long = latitude, lat = longitude)
pal <- colorNumeric(
palette = "viridis",
domain = data_2018$persons_injured)
data_2018 %>%
filter(!(lat < "-70" | lat >= "-75")) %>%
filter(persons_injured > 2) %>%
mutate(
label = str_c("<b>vehicle type: ", vehicle_type, "</b><br>Month: ", month , sep = "") ) %>%
sample_n(2000) %>%
leaflet() %>%
addTiles() %>%
addProviderTiles(providers$CartoDB.Positron) %>%
addLegend("bottomright", pal = pal, values = ~persons_injured,
title = "Persons Injured",
opacity = 1
) %>%
addCircleMarkers(
~lat, ~long,
color = ~pal(persons_injured),
radius = 0.5,
popup = ~ label)
data_2018 %>%
group_by(borough) %>%
summarise(n())
## # A tibble: 6 x 2
## borough `n()`
## <chr> <int>
## 1 BRONX 22121
## 2 BROOKLYN 46314
## 3 MANHATTAN 29728
## 4 QUEENS 40400
## 5 STATEN ISLAND 5988
## 6 <NA> 71583
data_2018_seperate = tidy_data
data_kill_injured = data_2018_seperate %>%
select(month, persons_injured,persons_killed, pedestrians_injured, pedestrians_killed, cyclist_injured, cyclist_killed, motorist_injured, motorist_killed)
data_kill_injured$injured = apply(data_kill_injured[,c(2,4,6,8)],1,sum,na.rm=T)
data_kill_injured$killed = apply(data_kill_injured[,c(3,5,7,9)],1,sum,na.rm=T)
data_kill_injured =
data_kill_injured %>%
group_by(month) %>%
summarise(
sum_injured = sum(injured),
sum_killed = sum(killed)
) %>%
ungroup()
data_kill_injured = data_kill_injured %>%
pivot_longer(
sum_injured:sum_killed,
names_to = "type",
values_to = "number"
)
plot_kill_injured = data_kill_injured %>%
ggplot(aes(x = month, y = number, color = type))+
geom_point()+
geom_line()+
scale_x_continuous(breaks=seq(1, 12, 1),
labels = c("Jan", "Feb", "Mar", "Apr", "May",
"Jun","Jul", "Aug", "Sep", "Oct",
"Nov", "Dec"))+
scale_y_continuous(breaks = seq(0,11000,1000))+
labs(
title = "Trend of People being Injured or Killed through the Year")+
theme(axis.title = element_text(size=14,face="bold"),
plot.title = element_text(hjust = 0.5,color ="Blue"))
plot_kill_injured = ggplotly(plot_kill_injured)
plot_kill_injured
data_kill_injured_day = data_2018_seperate %>%
select(date_complete, persons_injured,persons_killed, pedestrians_injured, pedestrians_killed, cyclist_injured, cyclist_killed, motorist_injured, motorist_killed)
data_kill_injured_day$injured = apply(data_kill_injured_day[,c(2,4,6,8)],1,sum,na.rm=T)
data_kill_injured_day$killed = apply(data_kill_injured_day[,c(3,5,7,9)],1,sum,na.rm=T)
data_kill_injured_day =
data_kill_injured_day%>%
group_by(date_complete) %>%
summarise(
sum_injured = sum(injured),
sum_killed = sum(killed)
) %>%
ungroup()
data_kill_injured_day = data_kill_injured_day %>%
pivot_longer(
sum_injured:sum_killed,
names_to = "type",
values_to = "number"
) %>%
mutate(
day = rep(1:365, each = 2),
month = rep(1:12,c(62,56,62,60,62,60,62,62,60,62,60,62))
)
plot_kill_injured_day = data_kill_injured_day %>%
ggplot(aes(x = date_complete, y = number, group = type, color = type))+
geom_line()+
theme(axis.text.x = element_blank(),
axis.title = element_text(size=14,face="bold"),
plot.title = element_text(hjust = 0.5,color ="Blue"))+
labs(
title = "Trend of People being Injured or Killed through the Day",
x = "Day of the Year")
plot_kill_injured_day

plot_kill_injured_day = ggplotly(plot_kill_injured_day)
plot_kill_injured_day
proportion of accident and injured people by borough, by hour
boro_plot = tidy_data %>%
drop_na(borough, persons_injured, persons_killed) %>%
group_by(borough) %>%
summarize(accident = n(),
injured = sum(persons_injured)) %>%
mutate(prop_accident = round(accident/sum(accident),4),
prop_injured = round(injured/sum(injured),4)) %>%
mutate(text_label1 = str_c("accident: ", accident),
text_label2 = str_c("injured: ", injured)) %>%
plot_ly(x = ~borough, y = ~prop_accident, type = 'bar', name = 'accident', text = ~text_label1, alpha = 0.8) %>%
add_trace(x = ~borough, y = ~prop_injured, type = 'bar', name = 'injury', text = ~text_label2, alpha = 0.8) %>%
layout(title="proportion of accident and injured people by borough", xaxis = list(title = 'borough'), barmode = 'group', legend = list(orientation = "h", xanchor = "center", x = 0.5, y = -0.2))
boro_plot
Number of accident and injured people in Brooklyn is greatest comparing with other boroughs. Queen is second. Staten Island have fewest accident and injured people. Manhattan and Bronx is in the middle.
hour_table = tidy_data %>%
drop_na(hour, persons_injured, persons_killed) %>%
group_by(hour) %>%
summarize(accident = n(),
injured = sum(persons_injured))
hour_plot = hour_table %>%
mutate(prop_accident = round(accident/sum(accident),4),
prop_injured = round(injured/sum(injured),4)) %>%
mutate(text_label1 = str_c("accident: ", accident),
text_label2 = str_c("injured: ", injured)) %>%
plot_ly(x = ~hour, y = ~prop_accident, type = 'scatter', mode = 'lines', name = 'accident', text = ~text_label1, alpha = 0.8) %>%
add_trace(x = ~hour, y = ~prop_injured, type = 'scatter', mode = 'lines', name = 'injury', text = ~text_label2, alpha = 0.8) %>%
layout(title="proportion of accident and injured people by hour", xaxis = list(autotick = FALSE, ticks = "outside", tick0 = 0, dtick = 1, title = 'Hour'), yaxis = list(autotick = FALSE, ticks = "outside", tick0 = 0, dtick = 0.01, title = 'proportion'), legend = list(orientation = "h", xanchor = "center", x = 0.5, y = -0.2))
hour_plot %>%
layout(shapes = list(type = "rect", fillcolor = "pink", line = list(color = "pink"), opacity = 0.3, x0 = 8, x1 = 19, xref = "x", y0 = 0, y1 = 0.08, yref = "y"))
Frequency of accident and injured people in the period of 8 am-19 pm is higher than other period of time. Started from 3 am, number of accident and injured people increase and reach a small peak at 8am. At 5pm, it reaches a big peak. then it started to decrease.
Holiday Analysis
data_2018 = tidy_data
holiday_data =
data_2018 %>%
select(date_complete, everything()) %>%
mutate(holiday = "No Holiday") %>%
select(date_complete, holiday, everything()) %>%
mutate(holiday = replace(holiday, date_complete == "01/01/2018", "New Year's Day"),
holiday = replace(holiday, date_complete == "01/15/2018", "Martin Luther King, Jr. Day"),
holiday = replace(holiday, date_complete == "01/13/2018"|date_complete =="01/14/2018", "Martin Luther King Day Weekend (2 days)"),
holiday = replace(holiday, date_complete == "02/19/2018", "Presidents Day"),
holiday = replace(holiday, date_complete == "02/17/2018"|date_complete =="02/18/2018", "Presidents Day Weekend (2 days)"),
holiday = replace(holiday, date_complete == "05/28/2018", "Memorial Day"),
holiday = replace(holiday, date_complete == "05/26/2018"|date_complete =="05/27/2018", "Memorial Day Weekend (2 days)"),
holiday = replace(holiday, date_complete == "07/04/2018", "Independence Day"),
holiday = replace(holiday, date_complete == "07/07/2018"|date_complete =="07/08/2018", "Independence Day Weekend (2 days)"),
holiday = replace(holiday, date_complete == "09/03/2018", "Labor Day"),
holiday = replace(holiday, date_complete == "09/01/2018"|date_complete =="09/02/2018", "Labor Day Weekend (2 days)"),
holiday = replace(holiday, date_complete == "11/22/2018", "Thanksgiving Day"),
holiday = replace(holiday, date_complete == "11/23/2018"|date_complete =="11/24/2018"|date_complete =="11/25/2018", "Thanksgiving Weekend (3 days)"),
holiday = replace(holiday, date_complete == "12/25/2018", "Christmas Day"),
holiday = replace(holiday, date_complete == "12/26/2018"|date_complete =="12/27/2018", "Christmas Day Weekend (2 days)")
)
holiday_data =
holiday_data %>%
drop_na(persons_injured)
holiday_data %>%
group_by(holiday) %>%
summarise(total_injured = sum(persons_injured))
## # A tibble: 16 x 2
## holiday total_injured
## <chr> <dbl>
## 1 Christmas Day 93
## 2 Christmas Day Weekend (2 days) 286
## 3 Independence Day 130
## 4 Independence Day Weekend (2 days) 330
## 5 Labor Day 216
## 6 Labor Day Weekend (2 days) 333
## 7 Martin Luther King Day Weekend (2 days) 279
## 8 Martin Luther King, Jr. Day 151
## 9 Memorial Day 156
## 10 Memorial Day Weekend (2 days) 357
## 11 New Year's Day 122
## 12 No Holiday 54633
## 13 Presidents Day 103
## 14 Presidents Day Weekend (2 days) 276
## 15 Thanksgiving Day 109
## 16 Thanksgiving Weekend (3 days) 382
bronx =
holiday_data %>%
filter(borough == "BRONX") %>%
group_by(holiday) %>%
summarise(bronx_total_injured = sum(persons_injured))
brooklyn =
holiday_data %>%
filter(borough == "BROOKLYN") %>%
group_by(holiday) %>%
summarise(brooklyn_total_injured = sum(persons_injured)) %>%
select(brooklyn_total_injured)
manhattan =
holiday_data %>%
filter(borough == "MANHATTAN") %>%
group_by(holiday) %>%
summarise(manhattan_total_injured = sum(persons_injured))%>%
select(manhattan_total_injured)
queens =
holiday_data %>%
filter(borough == "QUEENS") %>%
group_by(holiday) %>%
summarise(queens_total_injured = sum(persons_injured))%>%
select(queens_total_injured)
staten_island =
holiday_data %>%
filter(borough == "STATEN ISLAND") %>%
group_by(holiday) %>%
summarise(staten_island_total_injured = sum(persons_injured))%>%
select(staten_island_total_injured)
holiday_borough =
cbind(bronx, brooklyn, manhattan, queens, staten_island) %>%
mutate(total = bronx_total_injured+brooklyn_total_injured+manhattan_total_injured+ queens_total_injured+staten_island_total_injured) %>%
filter(holiday!="No Holiday")
knitr::kable(holiday_borough)
| Christmas Day |
10 |
25 |
5 |
24 |
6 |
70 |
| Christmas Day Weekend (2 days) |
34 |
99 |
23 |
41 |
3 |
200 |
| Independence Day |
18 |
32 |
3 |
18 |
1 |
72 |
| Independence Day Weekend (2 days) |
26 |
60 |
38 |
54 |
6 |
184 |
| Labor Day |
3 |
60 |
13 |
32 |
3 |
111 |
| Labor Day Weekend (2 days) |
31 |
89 |
10 |
62 |
7 |
199 |
| Martin Luther King Day Weekend (2 days) |
23 |
58 |
40 |
53 |
7 |
181 |
| Martin Luther King, Jr. Day |
27 |
42 |
14 |
19 |
4 |
106 |
| Memorial Day |
16 |
38 |
12 |
24 |
3 |
93 |
| Memorial Day Weekend (2 days) |
36 |
70 |
27 |
67 |
7 |
207 |
| New Year’s Day |
10 |
34 |
9 |
14 |
1 |
68 |
| Presidents Day |
5 |
32 |
8 |
32 |
1 |
78 |
| Presidents Day Weekend (2 days) |
30 |
42 |
28 |
51 |
4 |
155 |
| Thanksgiving Day |
13 |
17 |
5 |
17 |
0 |
52 |
| Thanksgiving Weekend (3 days) |
42 |
74 |
39 |
93 |
4 |
252 |
holiday_borough_day =
holiday_borough %>%
filter(holiday =="Christmas Day"|holiday =="Independence Day"|holiday =="Labor Day"|holiday =="Martin Luther King, Jr. Day"|holiday =="Memorial Day"|holiday =="New Year’s Day"|holiday =="Thanksgiving Day"|holiday =="Presidents Day")
knitr::kable(holiday_borough_day)
| Christmas Day |
10 |
25 |
5 |
24 |
6 |
70 |
| Independence Day |
18 |
32 |
3 |
18 |
1 |
72 |
| Labor Day |
3 |
60 |
13 |
32 |
3 |
111 |
| Martin Luther King, Jr. Day |
27 |
42 |
14 |
19 |
4 |
106 |
| Memorial Day |
16 |
38 |
12 |
24 |
3 |
93 |
| Presidents Day |
5 |
32 |
8 |
32 |
1 |
78 |
| Thanksgiving Day |
13 |
17 |
5 |
17 |
0 |
52 |
holiday_borough_day = holiday_borough_day[order(holiday_borough_day$total),]
# On the day of the holiday
plot_day =
pivot_longer(
holiday_borough_day,
bronx_total_injured:staten_island_total_injured,
names_to = "borough",
values_to = "person_injured") %>%
ggplot(aes(x = reorder(holiday, -total), y = person_injured, fill = borough)) +
geom_bar(stat="identity") +
theme(legend.position = "right", axis.text.x = element_text(angle = 45)) +
labs(
title = "Persons Injured by Car Crashs on Federal Holidays (2018)",
x = "Holiday",
y = "# Persons Injured")
# scale_fill_discrete(name = "Borough", labels = c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island"))
On three day weekends
data_2018 = tidy_data
holiday_data_weekend =
data_2018 %>%
select(date_complete, everything()) %>%
mutate(holiday = "No Holiday") %>%
select(date_complete, holiday, everything()) %>%
mutate(holiday = replace(holiday, date_complete == "01/15/2018"|date_complete == "01/13/2018"|date_complete =="01/14/2018", "Martin Luther King Day Weekend (3 days)"),
holiday = replace(holiday, date_complete == "02/17/2018"|date_complete == "02/19/2018"|date_complete =="02/18/2018", "Presidents Day Weekend (3 days)"),
holiday = replace(holiday, date_complete == "05/28/2018"|date_complete == "05/26/2018"|date_complete =="05/27/2018", "Memorial Day Weekend (3 days)"),
holiday = replace(holiday, date_complete == "07/04/2018"|date_complete == "07/07/2018"|date_complete =="07/08/2018", "Independence Day Weekend (3 days)"),
holiday = replace(holiday, date_complete == "09/03/2018"|date_complete == "09/01/2018"|date_complete =="09/02/2018", "Labor Day Weekend (3 days)"),
holiday = replace(holiday, date_complete == "11/22/2018"|date_complete == "11/23/2018"|date_complete =="11/24/2018"|date_complete =="11/25/2018", "Thanksgiving Weekend (4 days)"),
holiday = replace(holiday, date_complete == "12/25/2018"|date_complete == "12/26/2018"|date_complete =="12/27/2018", "Christmas Day Weekend (3 days)")
)
holiday_data_weekend =
holiday_data_weekend %>%
drop_na(persons_injured)
holiday_data_weekend %>%
group_by(holiday) %>%
summarise(total_injured = sum(persons_injured))
## # A tibble: 8 x 2
## holiday total_injured
## <chr> <dbl>
## 1 Christmas Day Weekend (3 days) 379
## 2 Independence Day Weekend (3 days) 460
## 3 Labor Day Weekend (3 days) 549
## 4 Martin Luther King Day Weekend (3 days) 430
## 5 Memorial Day Weekend (3 days) 513
## 6 No Holiday 54755
## 7 Presidents Day Weekend (3 days) 379
## 8 Thanksgiving Weekend (4 days) 491
bronx =
holiday_data_weekend %>%
filter(borough == "BRONX") %>%
group_by(holiday) %>%
summarise(bronx_total_injured = sum(persons_injured))
brooklyn =
holiday_data_weekend %>%
filter(borough == "BROOKLYN") %>%
group_by(holiday) %>%
summarise(brooklyn_total_injured = sum(persons_injured)) %>%
select(brooklyn_total_injured)
manhattan =
holiday_data_weekend %>%
filter(borough == "MANHATTAN") %>%
group_by(holiday) %>%
summarise(manhattan_total_injured = sum(persons_injured))%>%
select(manhattan_total_injured)
queens =
holiday_data_weekend %>%
filter(borough == "QUEENS") %>%
group_by(holiday) %>%
summarise(queens_total_injured = sum(persons_injured))%>%
select(queens_total_injured)
staten_island =
holiday_data_weekend %>%
filter(borough == "STATEN ISLAND") %>%
group_by(holiday) %>%
summarise(staten_island_total_injured = sum(persons_injured))%>%
select(staten_island_total_injured)
holiday_borough_weekend =
cbind(bronx, brooklyn, manhattan, queens, staten_island) %>%
mutate(total = bronx_total_injured+brooklyn_total_injured+manhattan_total_injured+ queens_total_injured+staten_island_total_injured) %>%
filter(holiday!="No Holiday")
knitr::kable(holiday_borough_weekend)
| Christmas Day Weekend (3 days) |
44 |
124 |
28 |
65 |
9 |
270 |
| Independence Day Weekend (3 days) |
44 |
92 |
41 |
72 |
7 |
256 |
| Labor Day Weekend (3 days) |
34 |
149 |
23 |
94 |
10 |
310 |
| Martin Luther King Day Weekend (3 days) |
50 |
100 |
54 |
72 |
11 |
287 |
| Memorial Day Weekend (3 days) |
52 |
108 |
39 |
91 |
10 |
300 |
| Presidents Day Weekend (3 days) |
35 |
74 |
36 |
83 |
5 |
233 |
| Thanksgiving Weekend (4 days) |
55 |
91 |
44 |
110 |
4 |
304 |
plot_weekend =
pivot_longer(
holiday_borough_weekend,
bronx_total_injured:staten_island_total_injured,
names_to = "borough",
values_to = "person_injured") %>%
ggplot(aes(x = reorder(holiday, -total), y = person_injured, fill = borough)) +
geom_bar(stat="identity") +
theme(legend.position = "right", axis.text.x = element_text(angle = 45)) +
labs(
title = "Persons Injured by Car Crashs on Federal Holidays Weekends (2018)",
x = "Holiday Weekends",
y = "# Persons Injured") +
scale_fill_discrete(name = "Borough", labels = c("Bronx", "Brooklyn", "Manhattan", "Queens", "Staten Island"))
plot_day + plot_weekend
